A diagnosis standardization method and device based on word vectors

Abstract

The invention discloses a diagnosis standardization method and device based on word vectors, and the method comprises the steps: S1, obtaining a to-be-processed diagnosis statement, carrying out the word segmentation of the to-be-processed diagnosis statement, and obtaining a word segmentation result; S2, according to the word segmentation result and a pre-constructed corresponding model, establishing a word vector, a word information amount and a part-of-speech tagging result of the to-be-processed diagnosis statement; S3, according to the word vector, the word information amount and the part-of-speech tagging result, respectively calculating the similarity between the to-be-processed diagnosis statement and a related standard diagnosis statement in a standard library; and S4, selecting astandard diagnosis statement most similar to the to-be-processed diagnosis statement as a standardization result of the diagnosis statement. According to the method, the semantic similarity between the to-be-processed diagnosis statement and the standard diagnosis statement closest to the current diagnosis statement is calculated, the standard diagnosis statement with the highest similarity to the to-be-processed diagnosis statement is selected as the standardization result, the corresponding model is updated regularly, and the diagnosis standardization accuracy is improved.

Description

technical field [0001] The invention relates to the technical field of data processing, in particular to a diagnostic standardization method and device based on word vectors. Background technique [0002] The standardization of disease diagnoses has important implications for insurance claims and medical research statistics. Current disease standards include ICD-10 (International Classification of Diseases) issued by international authorities. However, on the one hand, different hospitals have different ICD-10 diagnostic criteria, and on the other hand, there is a certain gap between the doctor’s handwritten diagnosis and the standard diagnosis. How to standardize the non-standard diagnosis in different hospitals is a very realistic problem. question of meaning. [0003] The current diagnostic standardization method has the following problems: [0004] 1. The unsupervised natural language processing method is completely adopted, and the existing labeling resources are not...

AI Technical Summary

Problems solved by technology

[0006] In order to solve the problems of the prior art, the embodiment of the present invention provides a diagnostic standardization method and device based on word vectors, so as to overcome the reliance on manual revision and labeling in the prior art that consumes a lot of human resources and completely uses unsupervised natural language The processing method does not use the existing labeling resources, and the accuracy rate is low, etc.

Images